چكيده به لاتين
Due to the increasing number of elderly people in the world and especially in Iran, the study of diseases related to these ages, especially Alzheimer's disease (AD), and the provision of treatment and control methods are among the requirements of future studies. General methods for diagnosing AD are to study specific areas of the brain, such as the hippocampus, that are related to memory. Because these parts are the first areas of the brain that are affected by this disease and lose their normal function. Resting-state fMRI (rs-fMRI) is capable of measuring the spontaneous fluctuations of brain activity without any task; hence it is less sensitive to individual cognitive abilities. Therefore, in this study we propose novel methods using fMRI and T1-MRI to discriminate between healthy participants (HC), EMCI, AD and mild cognitive impairment (MCI), Also, diagnosis of disease progression from MCI to AD. Data were extracted from the publicly available databases of the Alzheimer’s disease neuroimaging initiative database (ADNI) and MIRIAD. That include MRI, fMRI and DTI data. After preprocessing, we applied graph theory to extract a collection of graph parameters. This data is then given to five different evolutionary algorithms (EA) to select an optimum subset of parameters. These selected parameters are subsequently given to classification algorithms to classify the data into two groups of HC and EMCI. Also, using three automatic brain segmentation methods, we extracted volumetric parameters from T1-MRI as input to the optimization algorithms. Finally, this study presents a unique three-year longitudinal T1-MRI and DTI based method for predicting the MCI-to-AD conversion using high efficacy multiobjective evolutionary algorithm as feature selection. The results show that the proposed methods are helpful for early AD diagnosis and an essential role of T1-MRI and DTI in detecting AD conversion in MCI patients. All algorithms achieved classification accuracy of 95%, These accuracy percentages are 10% to 15% higher than previously studied.
